Ecom Data Cooler Talk Ep 8 What is Customer Lifetime Value CLV (and how to calculate it simply)

Hey guys, what’s up, John from Segments
here. Welcome to the eighth episode of Ecom Data
Cooler Talk, where we talk about important topics in ecommerce and data, next to a watercooler. We started this channel to make data science
accessible and give ecommerce data powers back to the people. If you like the videos, hit that like button
and subscribe to our channel, it really means a lot to me. Today we are going look at something that
almost everyone has heard of and wants to know, but is still a little confusing and
vague, and that is Customer Lifetime Value, usually known as CLV or LTV. Before we get started, I want to follow up
with my key takeaways from Day 2 of Shopify Unite. If you want to hear about my takeaways from
Day 1, check out episode 7, what is customers data. Unite day 2 started off with CTO Jean-Michel
Lemieux’s keynote about Shopify’s technology investments and infrastructure improvements. Here are my three key takeaways: 1. It’s all about security — Shopify has something
like 3000 whitehats they use to find vulnerabilities in the system with dev stores so your data
stays safe. They’ve paid out over $1M in bounties, and
with single-sign-on rolling out, they want to keep your data “yours” 2. It’s all about globalization — Shopify is
building the global retail operating system and they are building 4 new HQs in the EU,
China, and Australia. They want to enable borderless commerce, and
with more local hubs mean faster connections and better support. Lemieux compares it to building the silk road
from 1500 years ago in the level of engineering required to make it all click. 3. It’s all about speed and scale — With something
like 820,000 merchants from large to small, it’s a monumental task to achieve stability. Invariably, with more lines of code being
written into the platform things slow down. To make sites load faster, Shopify is planning
to deploy new liquid rendering for themes which will be 7x faster, as well as using
webP format for images. He also talked about merchants doing 8000
orders per minute and it’s barely visible in their overall system volume, which is meant
to instill trust when Black Friday Cyber Monday comes around. Ok, let’s dive in. What is CLV? Put Simply, customer lifetime value is a predictive
metric that measures the total spending of a customer’s relationship with your business. In other words, you’re trying to find the
total area under a customer’s revenue curve over time. CLV is useful in understanding how much a
customer is worth, and factors into how much you can spend on acquiring each new customer,
measured by the cost of acquisition or CAC. If your CLV is $50 and you’re having to
spend $100 on acquiring a customer with FB ads, then you’re likely to lose $50 on every
customer. So then the goal is to increase CLV or decrease
CAC in order to achieve profitability. And I’ve kind of drawn this curve without
explaining how I got it, so here’s what you do. First, you want to line up everyone by the
time they sign-up, in months or weeks, and look at when each of their orders occurred
relative to their sign-up date. So if I joined on 1/1/2019, and I made a purchase
on 1/1, 3/1, and 6/1, then my purchases will line up like with an x in month 1, another
x in month 3, and another x in month 6. And another person might have two x’s in
month 1, 1 in month 4. Or maybe 0 in month 1, one in month 3, and
one in month 5, etc. Remember, each line represents a single customer,
and each x represents an order with actual revenue. If you add up hundreds of customers purchases
this way, you will end up with the curve we started with, with revenue on the y-axis which
is the sum of individual orders, and time in months of the customer’s relationship
tenure on the x-axis. Finally, what you want to do is look at the
total area under the revenue curve for each month, divided it by the total number of customers
over the entire period and you will have calculated a simple CLV for ecommerce. Ok, so far so good right? Well, yes and no. There are a few things we need to explore
further, which we will go over next. Why goes into CLV calculation? When you google Customer Lifetime Value, you’re
going to get back lots of different answers, with varying degrees of complexity. Here is a very common one that states CLV
is the expected revenue from a customer, call this customer value multiplied by the avg.
tenure of the customer relationship. The customer value is the product of the Avg. Order Value and the Purchase Frequency. This is similar to the approach I mentioned
earlier, where the first approach works without first defining the length of the analysis
period. There are two important things to keep in
mind here — first is that ecommerce is a non-contractual business, unlike your cable
TV subscriptions which typically is on a 2-year contract, online shoppers are free to come
and go at any point. Secondly, customers are free to purchase at
any point, not at fixed time points like prescription drugs. Therefore when a customer stops purchasing,
there is a degree of uncertainty whether this person is going to come back and buy again
or not. Remember our previous chart which kind of
ended at 6 months? I didn’t exactly specify it, but let’s
say this is a new store with just 6 months of data, then what’s to the left is what
actually happened, let’s call this existing equity. Then what’s to the right is the unknown,
and let’s call this future equity. The way most of the CLV calculations differ
is in how we decide to treat the uncertainty in the future. In the simple case, we will make the assumption
that future purchases behave similarly to existing purchases and project forward. In the more complicated, more accurate approach,
which I will call the Buy-until-you-die model-based approach, is to look at customers purchase
patterns in terms of the frequency of purchase and how much time has gone since their last
purchase. When you are just starting, you won’t have
a lot of data to get good results with the model-based approach. Instead, you can opt for the simple approach
when you have around 6 months of stable sales with repeat orders. Once you have enough data and orders, and
we usually recommend 1 to 2 years of order history with a few thousand orders, then you
can utilize the model CLV which will likely give you better estimates overall. Ok so we’ve covered the major differences
in CLV and a few examples, you should be well on your way to discover CLV for your business. If you want to get started right away, you
can also get your CLV with Segments, where it’s available out of the box. Next, let’s look at a few example use cases
with CLV and why you should care. What can CLV do for me? CLV is an elusive metric and sometimes hailed
as the holy grail of measuring ROI. While there are many useful applications,
here are three examples to get you started. #1 CLV as a KPI: Now that we have CLV, we
should be constantly measuring and trying to improve it. As we discussed earlier, CLV is made up of
avg. order value, purchase frequency, and avg. tenure. Therefore, you can undertake initiatives in
each of those areas to improve CLV. Customer retention is usually the most important,
as long as the customer continues spending with you their CLV will likely rise exponentially. With Segments, we see that Loyal customers
with 3 or more purchases often have 7x to 10x higher CLV than one time purchasers. You should also be looking at how CLV is trending
each month to see if your efforts end up moving the needle on CLV month over month. This is also available in Segments. #2 CLV by channel and Subscription CLV: You
can calculate your CLV by acquisition channel and look at which channel is the most profitable
with the highest CLV, whether they found you by search, on facebook or Instagram, or email
campaigns. This can help you align your resources on
the best channel for your store. Another common use case we see is to understand
the value of adding a subscription to your store. We’ve analyzed CLV for Subscription vs.
regular customers and found them to be at least 3x to 5x more valuable than non-subscription
customers, even with discounted pricing on subscriptions. #3 CLV for marketing and support: Start measuring
CLV for different segments of your business, and prioritize accordingly. For example with RFM, you can readily see
how valuable each segment is, and for those high CLV customers who are about to churn,
you would be willing to offer more incentives in order to keep their business. This also applies to customer support, where
you could tailor the urgency and level of response according to how valuable each customer
is. If your MVP customer with a 10x higher CLV
is having issues with an order, you may want to jump in personally to make sure it gets
resolved quickly. In summary, we looked at what is CLV, what
goes into the calculation of CLV along with a few watch-outs, as well as useful applications
of CLV to find profitable channels and prioritizing marketing efforts. If you have more questions about CLV you can
always find us on Facebook, LinkedIn or through our app. Keep in mind that CLV is a predictive metric
meant to approximate the total potential value of a customer. As with all approximations, it is never exact
and changes based on the different business settings. For example, maybe your business has grown
so much that the CLV of today varies drastically from 2 years ago. However, As long as you’ve followed the
guidelines and practice common sense in your reasoning, CLV can be very useful to understand,
measure, and prioritize areas of your business. Finally, data is power and we want to give
ecommerce data powers back to the people. So join us and start growing with your data
today! If you enjoyed the video, don’t forget to
hit subscribe!

Leave a Reply